Fast indoor localization for exhibition venues with calibrating heterogeneous mobile devices

Indoor localization has attracted much attention these years and been applied into different fields and sites. This paper focuses on Wi-Fi based localization system in exhibition venues which have some unique features, for example, many heterogeneous mobile devices with different transmission characteristics. We propose a customized indoor localization system and an efficient positioning approach which can give fast and accurate location estimations. For tackling heterogeneous mobile device problem, we use Expectation-maximization (EM) algorithm to infer the parameters of the signal-strength-to-distance model so that each mobile device has its own signal strength degradation model. This is suitable for the exhibition scenario since it not only improves the localization accuracy for diverse mobile devices, but also reduces the offline data acquisition time as in the case of employing the fingerprinting approach. We have conducted experiments in our testbed as well as pilot tests at the Hong Kong Convention and Exhibition Centre (HKCEC). The experimental results show that our system has good localization performance and is robust to heterogeneous mobile devices.